The Research Of Network Partitioning For Secondary Voltage Control Considering Grid-Connected Wind Farm

Modern power systems are characteristic of large scales and complicatedstructures, into which generation of new energy forms like wind farms have beenintegrated in recent years, and the traditional voltage control either in a centralizedway or in a distributed way has been challenged to some extent. Generation power ofwind farms fluctuates randomly and continuously, and most of them need largeamounts of inductive power for excitation. Consequently, more and more seriousbecomes the problem of voltage stability of power systems with large-scale windenergy generation incorporated. Under these circumstances, it is necessary to partitionthe power system into a certain number of areas in order to monitor, analyze andcontrol the system effectively. Among all kinds of partitioning strategies, thehierarchical voltage control of Japan and France is the most successful. Thesecondary voltage control has a good effect on preventing power systems fromvoltage collapse, and leads to the reactive power optimization to some extent. Thepartitioning method is the prerequisite of the control strategy, and whether it isreasonable and efficient is of utmost importance.The existing partitioning methods have rarely involves cases in which windgeneration was incorporated, and studies with the randomness considered are muchmore difficult to be found. In this paper, the characteristics of wind generation realpower output are taken into consideration, and methods of probability statistics areused to build the scenarios by the probabilities of different wind generation realpower output. The expected scenarios are taken into partitioning process; thecomprehensive voltage/var sensitivity matrix is set up, in which the information ofelectric source buses and load nodes; and then, based on cluster and mappingalgorithms, a partitioning method of secondary voltage control is proposed, whichtakes the wind power generation into account and has the relative advantage over themethods without considering wind generation. The simulation of IEEE39buses verifies the reasonableness and effectiveness ofthe proposed method. By introducing cluster and mapping algorithms, the method isadjusted to partitioning power systems with large-scale wind power generationincorporated for secondary voltage control; by introducing expected scenarios ofdifferent wind generation output, the problem of comprising between the outputrandomness and the required certainty of partitioning results is solved to a largeextent; the fitness calculating and perturbation analysis of partitioning results showthat the proposed method is of sufficient flexibility to accommodate a certain amountof fluctuating of wind power, and perturbations caused by system faults.